▎ 摘　　要

An experimental probabilistic approach for health risk assessment was applied for graphene nanoplatelets (GNPs). The hazard assessment indicated a low level of toxicity for the GNPs. The benchmark dose method, based on sub-chronic and chronic inhalation exposure studies, was used to quantify a guidance value (BMCh) for occupational inhalation exposure to GNPs, expressed as a lognormal distribution with a geometric mean geometric standard deviation of 0.212 +/- 7.79 mg/m(3) and 9.37 x 10(4) +/- 7.6 particle/cm(3). Exposure scenarios (ES) were defined based on the scientific literature for large-scale production (ES1) and manufacturing (ES2) of GNPs; a third ES, concerning in-lab handling of GNPs (ES3) was based on results of experiments performed for this study. A probability distribution function was then assumed for each ES. The risk magnitude was calculated using a risk characterization ratio (RCR), defined as the ratio of the exposure distributions and the BMCh distribution. All three ES resulted in RCR distributions >= 1 (i.e. risk present); however, none of the ES had a statistically significant level of risk at a 95% confidence interval. A sensitivity analysis indicated that similar to 75% of the variation in the RCR distributions was due to uncertainties in the BMCh calculation.